- Published on
Sony and Warner sue Anthropic over 'one of the largest and most blatant ongoing thefts of intellectual property in hi...
- Authors
- Name
- DeepHorus
Major music publishers Sony Music and Warner Music have filed a federal lawsuit against artificial intelligence company Anthropic, describing the case as "one of the largest and most blatant ongoing thefts of intellectual property in history." The complaint, filed in Northern California, targets Anthropic and its executives personally, including CEO Dario Amodei and co-founder Benjamin Mann.
The lawsuit alleges Anthropic illegally used tens of thousands of copyrighted musical compositions, primarily song lyrics, to train its Claude AI models without permission. The plaintiffs claim the company acquired this data through unlawful means, specifically via torrenting from pirate libraries and unauthorized scraping from licensed platforms.
The complaint details that Anthropic allegedly torrented at least seven million books from pirate sources LibGen and PiLiMi. It also accuses the company of scraping song lyrics from platforms like MusixMatch and LyricFind in violation of their terms of service. Furthermore, Anthropic is alleged to have scanned and destroyed used songbooks and sheet music collections.
This legal action follows a precedent set in September 2025, when Anthropic agreed to a $1.5 billion settlement with authors and publishers for using pirated books. That case established that acquiring training data through illegal downloads constitutes a standalone copyright infringement, a strategy the plaintiffs are now replicating.
The plaintiffs are seeking significant damages, including up to $150,000 per infringed work and up to $25,000 per violation for removing copyright management information. They argue that even if Anthropic did not use the copyrighted works directly in commercial models, the illegal acquisition itself is the core violation.
The lawsuit also challenges Anthropic's use of common AI training datasets like Books3, The Pile, and Common Crawl, which allegedly contain unauthorized content. Additionally, it accuses the company of using a non-commercial model trained on pirated data to generate synthetic information, which was then used to train commercial Claude models.
In a related development, a November 2025 ruling by the Munich Regional Court established that copyright-protected lyrics within a model's parameters count as reproductions. This ruling holds model operators responsible for what their models produce, further strengthening the legal landscape for copyright holders against AI companies.
The case is expected to advance into a discovery phase that will uncover the full scope of Anthropic's data acquisition and training methods. The outcome could set a major precedent for how AI companies source and utilize copyrighted materials for model development in the future.
